We have recently switched our 1972 GE Frame 5 gas turbine from natural gas to liquid fuel only and cannot get it to light off. It is pre Atomising Air compressor model and was originally diesel only.

What I am looking for is a chart or graph relating rotor speed to compressor discharge pressure as we have looked at all the other components fuel, fuel pressure, fuel volume, spray pattern, fuel nozzle's replaced, ignitors checked, ignitors replaced time and again.

At crank (GT rotor at 850RPM) we get a maximum CDP of 5 kPa (0.72Psi) and I believe that's a bit low but cannot find any documentation to support this theory.

Thanks
Chris W

 

ChrisW,

This sounds a lot like another current thread:

http://control.com/thread/1434461126

Axial compressor discharge pressure is very low at low speed. It really doesn't start to increase very much above even 1 barg until shaft speed gets above about 80% or thereabouts. And, after that it starts increasing very fast (exponentially). This is characteristic of axial compressors--they're very inefficient at less than rated speed, but at rated speed they're very efficient.

You say the ignitors have repeatedly been replaced. What were the conditions of the ignitors when they were removed immediately after firing attempts? Were they wet or dry?

What does the exhaust stack look like during and just after firing? Is there any white vapour visible in the exhaust? That would indicate unburnt, atomized fuel. It should only be a faint white vapour, usually. No white vapour, or excessive white vapour would signs of problems. No white vapour would probably mean very little atomization, and in this case there should be a lot of fuel coming out of the false start drains (I'm presuming the unit has false start drains, on the lower combustion cans, and the lower compressor discharge casing, and usually also the bottom of the exhaust plenum).

Excessive white vapour would usually indicate too much fuel--which would also show up as extremely wet ignitor tips. Too much fuel can actually "drown" the spark, which is somewhat contrary to popular belief (which is usually more fuel is better). As you've noted, it's usually poor atomization that's the problem--and with mechanical atomization it's probably even more difficult.

Please monitor the exhaust stack discharge during and after the firing attempts--and the discharge coming from the false start drains (again--presuming the unit has them, and I believe most all liquid fuel fired units should have them--whether they are manual or "automatic" (driven by compressor discharge pressure).

If the GT exhausts into an HRSG ("boiler") it may take a little longer for any white vapour to be visible. If it's particularly cold where the rig is located, it might also be condensing in the HRSG--which ain't good, either, because when a flame is finally established it could be disastrous.

Unfortunately, not being able to see the drawings of the liquid fuel nozzles and mechanical atomizers and not having personal experience with these types of liquid fuel nozzles/atomization method it's very difficult to say what the problem may be. I presume the mechanical atomizer has some moving part to help with atomization? Do they move freely?

The liquid fuel nozzles I'm familiar with have an internal set of passages that imparts a swirling motion to the liquid fuel is it leaves the end of the nozzle--but there is no moving part. It's just a set of passages internal to the nozzle that are tangential to the axis of the nozzle. The fuel pressure does the "work" of imparting the swirling motion and as the fuel leaves the tip of the nozzle it is partially atomized, and mixes with atomizing air. I know you don't have atomizing air, but maybe some of this is pertinent.

Are you sure of the liquid fuel check valve cracking pressures and the high-pressure liquid fuel pump discharge pressure? Is the gauge on the liquid fuel flow divider recently calibrated and working properly? Some older machines actually used two check valves in series to develop the required pressure for atomization (two 50 psig check valves were used in each fuel nozzle supply to achieve 100 psig pressure for proper pressure atomization).

Fuel, spark, and air--and for liquid fuel-fired GE-design heavy duty gas turbines the list of requirements for establishing flame includes proper atomization, however it's achieved. What "powers" the mechanical atomizers? Is it fuel pressure? Is it compressor discharge pressure (not likely--but, without being able to see drawings, gotta ask)? If it's fuel pressure that's "powering" the mechanical atomizers, is the fuel pressure adequate? How long does it take to build up sufficient fuel pressure to overcome check valve pressure? 10 seconds? 30 seconds? 50 seconds?

Again, if the ignitors are very wet then it sounds like either there's poor atomization or too much fuel--perhaps a combination of both. Too much fuel with "decent" atomization would likely show up as a heavy white vapour in the exhaust (if the ambient isn't too cold causing the atomized fuel to condense in the exhaust--we don't know very much about the ambient conditions at your location). Too much fuel with "decent" atomization would also show up as a high flow out of the false start drains.

Too little fuel would probably not result in much white vapour (depending, again, on ambient conditions) but it would not usually result in wet ignitor tips.

If the unit doesn't have trouble lighting off on gas fuel I doubt just changing the fuel nozzles would have any effect on compressor discharge pressure. It's most likely poor atomization, too much fuel, or too little fuel.

Hope this helps--sincerely wish I could offer more!

 

ChrisW,

One more possibility.... What kind of liquid fuel flow divider does the unit have? A long, "single shaft" arrangement of ten individual flow dividers all coupled together? These arrangements have been known to "break". If the speed pickups are both located at the same end of the flow divider, they could be seeing quite a different flow-rate than the other end of the shaft.

This would usually show up as two distinctly different sets of fuel pressures to the ten nozzles as measured by the selector valve/gauge on the discharge of the flow divider (I'm presuming there is a selector valve and gauge which can be used to monitor/check fuel nozzle pressures).

 

Wow thanks for the prompt response, I'll start with the last question first.

Our machine has a long single shaft divider with NO speed pick-ups, we uncoupled all 10 cans and ran the fuel into buckets the result was the same amount of fuel in each bucket. The pressure spread across the divider is within GE spec from the owners manual. We have had the fuel tested and it has come back saying its good Diesel (my comment to the others was that the starter motor starts and runs fine so the fuel should be ok). We have single in-line check relief valves that lift with nitrogen applied at approximately 100 Psia.

Chris W

 

Thanks I'll check out the other thread.

I'll try and answer these questions without missing any.

For a bit of background I'm relatively new to these machines a few weeks in the job and the last time I worked on this particular girl or any OCGT was 1984, I've been trained by GE on them but that was then. For the intervening years I've been looking after Parsons steam sets, so to say I'm a bit rusty is like looking at this machine.

Right here we go, on crank we have large volumes of White/Gray smoke/vapour pouring out of the stack and everywhere smells of diesel. The false start drain is piped into a closed system for environmental reasons (We are right on the beach) so we have not witnessed how much fuel is lost to that.

When the mechs pulled out the spark plugs I believe they were wet,the igniter voltage is 135 KV and on test crackles at the tip violently, however there is no visible ignition through the viewing port on the can and the two wheel space thermocouples in the underside of the casing now have diesel constantly dripping out of them.

The atomisation of fuel is achieved by the nozzles and pressure only there is no mechanical atomiser, we have tested them on the bench into a bucket and seemed to produce a reasonable spray pattern and we have swapped them for a new set. This machine is a unmanned remote start straight open cycle black start machine for system emergencies.

A few years ago (also on diesel) the exhaust stack exploded that made quite a mess. Regarding the fuel pressure rise I haven't been watching that side of the machine so I don't know how long it takes to reach the cracking pressure of the check valves.

We never had any problem lighting off on gas but we no longer have gas available.

Thanks again for your effort its much appreciated.
Chris W

 

Chris W,

Wow! Thanks for all the good information in response to all my questions! (I'm not accustomed to that....)

If you have heavy white smoke (as opposed to vapours) coming out of the stack, do you see any exhaust temperature increase during firing? Another cause of heavy white smoke is when some combustors actually light and the heat from that combustor/those combustors will start to combust the unburnt liquid fuel in the exhaust--and that will greatly increase the amount of white smoke--as opposed to white vapours.

White vapours are just unburnt, atomized liquid fuel. It's difficult to describe; if there's no exhaust temperature increase and there is heavy white vapours then it's likely there's too much fuel. If there's heavy white smoke, billowing, and quite often drops to the ground from the stack and you see some exhaust temperature increase (100-150 deg F, for example) then it's likely that one or two (or three) combustors are lighting but the flame is not propagating through the cross-fire tubes to the other combustors to the cans with flame detectors.

I've never been able to see the spark from the ignitors through the sight glasses; only flame. The sight glasses are just too small. Have you checked the flame detectors? There are usually isolation valves (gate valves) between the detector and the combustion can on older machines; have you checked the isolation valves? I've also seen the sight tubes (carbon steel) develop a lot of rust, which doesn't help with flame detection.

If you're seeing diesel coming out of the wheelspace temperature penetrations I'm leaning towards too much fuel. Especially if the ignitor tips are wet (more so if they're VERY wet) after a firing attempt, the fuel can just drown the spark.

If the exhaust temperature increases slightly during firing, there is flame in one or more combustors--which is encouraging. Too little fuel can cause the flame not to propagate through the cross-fire tubes, and so can too much fuel. But, from the sounds of things if it ever does "fire" it could be pretty disastrous.

If the unit is dual fuel (gas/distillate) it probably has liquid fuel purging, which includes liquid fuel purge check valves close to the fuel nozzle. These have a bad habit of developing leaks if the unit runs a lot on gas fuel. If they leak when running on liquid fuel, the liquid fuel will flow through the three-way valve (VA-19, I think) out through the Tell-Tale Leak-off. But, this should be accompanied by low fuel pressure to that combustor--which you are not reporting.

Please write back to let us know what you find!

 

Hi this is a short reply as I'm not back at work till Tuesday this week so my information is where we were up to on Friday afternoon. I believe the smoke/vapour is un-burnt fuel, it certainly smells like it and it tends to drop to the ground where we can smell it.

There is no sign through the windows of a flame in either of the cans with igniters. We have not stripped either of the flame detectors yet as we don't believe we are getting any flames. The exhaust thermocouples are seeing about a 2 or 3 degree Centigrade rise during start attempts this is evenly spread across the T/C's and what we expected for compression of the air through the compressor (oh our ambient daytime temp as its nearly mid winter gets up to about about 20 deg C by the way with a min of about 14/15 ).

We have checked the purge check valves and they are tight, they use fuel gas as a purge medium on this machine and as I mentioned we have no gas now so that line is shut and isolated.Thanks again for the reply I believe your thoughts on excessive fuel are currently the most likely option. Will keep you posted.

Chris W.

 

Chris W,
By the way, flame must begin in one of the two combustors with ignitors, and then spread around the machine through the cross-fire tubes. So, if one or two combustors are lighting off it's likely one or both the combustors with the ignitors, so the cans should be warmer than the other adjacent cans. (Just mentioning this as a possible way to confirm the heavy white smoke is the result of one or two combustors lighting off but the flame not spreading around the machine.)

Frame 5s don't usually require too much fuel flow to establish flame.

I'm also curious because you said you had recently switched the machine from gas to liquid. Did you remove any hardware, devices or piping from the machine? What was done to the nozzles for this switch--or do you just mean you selected liquid fuel instead of natural gas. Please clarify, explain or describe what was done as part of this switching from gas to liquid?

Thanks!

 

Chris W,

Even uncombusted (or partially combusted) diesel vapours smell distinctly like diesel.

What is the turbine control panel using for liquid fuel flow feedback? Is there an RVDT on the high-pressure liquid fuel pump swash plate axis/shaft? Is it calibrated correctly?

Sorry; I've no experience with these older units. But if you have lots of fuel vapors coming out of the stack and falling to the beach it sure sounds like too much fuel to me, too.

Most machines I've worked on may have wisps of white vapours coming out of the stack just as the unit ignites, or after a failed start attempt on liquid fuel, but not so much they are so heavy they fall to the ground in "clouds." Sometimes when flame is lost in one or more combustors during warm-up and acceleration there be clouds of white smoke (billowing clouds) that sometimes fall to the ground. It's very like a diesel lorry that has a cylinder or two with low or no compression and lots of white smoke coming out of the exhaust pipe.

But you say there's negligible exhaust temperature increase, but lots of vapours. Sure sounds like excessive fuel. Especially if the vapours start very shortly after liquid fuel starts flowing.

Thanks for the continued feedback and updates! We look forward to a successful resolution and hearing the outcome.

 

Hi and thanks for all the help.

We have managed to get the old girl away by various devious means. As far as the system fuel goes we just selected "liquid" with no other changes. We did however remove the gas connections to each of the cans and blanked of the flanges. "They" have assured us that we can get the gas supply back but that's never likely to happen in reality.

This machine has been retro fitted with the latest control system but without upgrading or adding any of the modern control equipment IE, Compressor discharge pressure tx, fuel forwarding divider speed sensors ect.

How we got it away in the end was by fitting "Hitachi frame 9" ignitors to her, this got a visible flame in the can (tripped on temperature spread first time) and then we trimmed the fuel flow until we got a solid start. Smoked like an old steam engine for about 20 minutes then we ran her for the rest of the day and did an MCR run (21.5 Mw) then shut her down hoping no one asks for a re start..

thanks again Chris.

 

Chris W,

Thanks for the feedback! It's always welcome, and very useful for many.

I'm going to guess, one, the igniter length was either shorter or longer than the original putting the spark out of the way of the liquid fuel "spray" keeping the tips from getting too wet.

And, my second guess is that for some reason the spark was stronger, possibly the original plug tips weren't gapped correctly, or, there may have been some short inside the igniter assembly causing the spark to be weak.

The smoking like a steam engine was likely due to all the residual fuel caught in low points, and also which had made its way into the seams of the plates in the exhaust. You're likely lucky there wasn't a bang shortly after lightoff--hopefully because you purged sufficiently and often.

Establishing flame during firing requires fuel, spark, air--and atomization for liquid fuel. All have to be in nearly proper proportions, but not exactly perfectly stoichiometric. And, the igniter tips have to stay dry and develop sufficient spark.

I'd also add that the control system was likely putting too much fuel into the machine originally.

I wonder if blanking off the fuel purge had much effect, given there was no fuel gas pressure/flow available. Also, I would imagine the purge wasn't active until later in the acceleration, possibly after FSNL.

Usually issues like these are the result of two or more things combining in a 'perfect storm' of sorts--making troubleshooting and resolution time-consuming, not to mention frustrating. But time, patience and perseverance solve most problems--along with some logical thinking.

At any rate, congratulations on your success! And thanks again for the feedback!!!